Analysis of variance in production quality among manufacturing lines using the ANOVA method: A case study in a longitudinal rolling mill in a steel bar factory

Abstract

The research attempts to conduct an extensive statistical test on the variance of the overall production quality between two lines in the bar mill longitudinal rolling process using Analysis of Variance (ANOVA), focusing on the three most relevant factors for quality assessment: yield rate values, rate of production deviation, and dimensional conformance. The study examines the values for the significance of the variations encountered in the three factors. After reorganizing the longitudinal data on four weeks' worth of actual production process data to the daily level, the number of observations for each line was n=28. This research analyzed the data utilizing MATLAB with the aid of one-way, two-way, and repeated-measures ANOVA models with the Bonferroni adjustment for the comparisons. The results showed that although the “Compliance” value did not present any significant differences among the two lines (F=0.041, p=0.841), it did present a significant difference with regard to the deviation rate (F=8.921, p=0.006), with Line2 deviating by 102%. The D25 performed better, and the “Diameter Effect on Yield” presented strong significance values (F=14.237, p=0.005). Recommendations for improvements in monitoring and further optimization for the process have been discussed in the final conclusions drawn for the research.